ING Direct 2015 Annual Report Download - page 259

Contents
Report of the
Executive Board
Corporate
Governance
Consolidated
annual accounts
Parent company
annual accounts
Other
information
Additional
information
Additional Pillar III information - continued
Regulatory Exposure at Default (READ): The second element is the counterparty’s exposure at default. These models are intended to
estimate the outstanding amount or obligation at the moment of default in the future. Since the fact that a counterparty will go into
default is not known, and the level of outstanding that may occur on that date is also not known, ING Bank uses a combination of
statistical, expert and hybrid models to estimate the Exposure at Default. With the exception of guarantees and letters of credit, the
EAD is always equal to or higher than the associated credit risk outstanding, under the assumption that counterparties tend to absorb
liquidity from available credit resources before financial problems become apparent to the counterparty’s creditors. The EAD is largely
a function of the type of credit facility (revolving, overdraft, term) offered to the borrower.
Loss Given Default (LGD): The third element is the loss given default. These models are intended to estimate the amount ING Bank will
lose when liquidating collateral pledged in association with a given loan or financial obligation, or alternatively, liquidating the
company as a whole, as part of a workout process. LGD models are based on cover types, estimated recovery rates given orderly
liquidation, and (in) direct cost of liquidation. For financial collateral for counterparty credit risk, ING Bank uses the Financial Collateral
Comprehensive Method to allow for mitigation effects.
Maturity (M): The fourth element is the time to the maturity of the underlying financial obligation. Regulations (CRR/CRD IV) cap the
maturity element at five years, despite the fact that many obligations extend their facilities for longer than five years.
Expected Loss (EL): The expected loss provides a measure of the value of the credit losses that ING Bank may reasonably expect to
incur on its portfolio. In its basic form, the expected loss can be represented as: EL = PD * EAD * LGD. ING Bank must maintain a capital
buffer against unexpected losses in order to protect itself against credit losses associated with unusual market events outside of the
statistical norms.
Exposure Class: The exposure class (a grouping of credit risks associated with a common obligor type or product type) is a driver for
the correlation factor. To calculate Capital the default correlation between a transaction and all other transactions in the portfolio are
taken into account. The correlation factor determines which portion of the standalone risk of a transaction is retained when the
transaction is included in the portfolio and the portfolio diversification benefits are taken into consideration.
AIRB models per exposure class
Within ING Bank internal Basel models are used to determine the PD, EAD and LGD for regulatory and economic capital. Bank wide, ING
Bank has implemented around 90 models, including various sub models for specific portfolios. A model may be applicable for various
exposure classes. In the table below, the number of significant PD, EAD, and LGD models per asset class are shown. Additionally a
description of the model and methodology are provided per exposure class. The asset classes presented in this table do not align with
the Basel Exposure classes as the scope has been redefined to better fit the scope of the model. SME exposure for example can be part
of both Corporate exposures as Other Retail depending on the size of the SME.
A
sset classes
measured
READ for
associated
Asset class
% of
Total
EAD
RWAs for
associated
Asset class
% of
Total
RWA
Model
Type
Number of
significant
models
Model description and methodology
Number
of years
of data
Sovereigns
107,835 13.6% 5,455 2.7%
PD
1
The Government Central PD model is a fully statistical
model, containing only quantitative risk drivers.
>7
years
LGD
1
The LGD model for Sovereigns and other governments is
an unsecured recovery model built on
assessment of
structural factors that
influence a country’s performance.
>7
years
EAD
1
The Low Default EAD model is a hybrid model that pools
default information from multiple low default portfolios,
including central governments and central banks.
>7
years
Government
related
entities
13,774 1.7% 2,356 1.2%
PD
1
The government related entities PD model is expert based
and assigns ratings based on stand
-alone credit
fundamentals as well as degree of government support.
>7
years
LGD
1
The LGD model for Government related entities is a
secured/unsecured recovery model built on assessment
of stand
-alone fundamentals as well as geography.
>7
years
EAD
1
The Low Default EAD model is a hybrid model that pools
default information from
multiple low default portfolios,
including government related entities.
>7
years
AIRB models and methodology
ING Bank Annual Report 2015 257